Understand this story
On October 29, 2026, Brown University assistant professor Randall Balestriero and Yann LeCun published a 37-page paper, "Proving and Improving the Platonic Representation Hypothesis with JEPAs," proposing that, under the condition that positive sample pairs are two observations of the same scene, JEPAs trained separately on any number of modalities will learn the same representation differing only by a rotation. The paper states that next-token prediction is a JEPA using bad views, which explains why language models can only rely on scale and vision alignment. On the experimental side, a 109-million-parameter text JEPA achieved 0.184 on the cross-modal representation alignment metric, higher than Qwen2.5-72B's 0.154 and Llama-3-70B's 0.151, but this metric measures the closeness of representation structure, not general capability. Regarding training sample efficiency, earlier reports stated that good views require 42 times fewer samples than next-token prediction, and subsequent reports stated that the gap is about 14 times when calculated by token count.
AI-generated from public reports
Latest development · Report
LeCun Team Paper: 109M-Parameter Text JEPA Outperforms 72B-Parameter Qwen2.5 in Cross-Modal AlignmentDevelopments
1 confirmed developments
Reports and discussion timeline
Browse reports and discussion by time. Their evidence levels remain distinct.
Nov 2
36Kr · AI
LeCun Team Paper: 109M-Parameter Text JEPA Outperforms 72B-Parameter Qwen2.5 in Cross-Modal Alignment
The paper by LeCun and Randall Balestriero uses the JEPA framework to provide a mathematical proof of the Platonic Representation Hypothesis, and claims that on the cross-modal representation alignment task, a 109M-parameter text JEPA outperforms the 72B-parameter Qwen2.5. The research shows that Good Views training requires only 21.6 million samples to reach the alignment level that a next-token prediction model achieves with 900 million samples, reducing the sample count to 1/42, or to 1/14 when counted in tokens. The authors also point out that this comparison is limited to specific alignment metrics, and that the mathematical proof relies on idealized conditions such as a finite world and reversible observations.
36Kr · AI
LeCun and Balestriero Publish JEPA Paper: 109-Million-Parameter Model Surpasses… in Cross-Modal Representation Alignment
Yann LeCun and Brown University assistant professor Randall Balestriero have published a 37-page paper, "Proving and Improving the Platonic Representation Hypothesis with JEPAs," proposing that under positive-sample relationships meeting specific conditions, JEPA models of different modalities can learn representation spaces that differ only by a rotation transformation. In experiments, a 109-million-parameter text JEPA scored 0.184 on the cross-modal representation alignment metric, higher than Qwen2.5-72B's 0.154 and Llama-3-70B's 0.151, but this metric measures the closeness of representation structures, not general capability. The research also claims that Good Views training reaches the alignment level of next-token prediction with 900 million samples using about 21.6 million samples, a gap of roughly 14 times when calculated by token count.
Nov 1
Hacker News Trending
Randall Balestriero and Yann LeCun use JEPA to prove and improve the Platonic representation hypothesis
Randall Balestriero and Yann LeCun propose the Platonic representation theorem: as long as positive sample pairs are two observations of the same scene (good views), JEPAs trained separately on any number of modalities will learn identical representations differing only by a rotation. They point out that next-token prediction is a JEPA using bad views, which explains why language models can only rely on scale and vision alignment; to achieve equivalent alignment on the same corpus, good views require 42 times fewer training samples than next-token prediction, and a 109M-parameter text JEPA aligns with DINOv2 and CLIP image encoders better than all evaluated language models of up to 72B parameters.
Heat trend for this story
- Current
- 15
- Period peak
- 16 Nov 2 15:00
- 24-hour change
- —
Move the pointer or select the chart to inspect hourly heat. Use the left and right arrow keys with a keyboard.